cis-2,2′-Bipyrimidine-Bridged Polynuclear Complex: A Stairway-like

Aug 31, 2010 - Reflectance spectroscopy did not allow one to distinguish an intervalence charge-transfer band. However, time-dependent density functio...
0 downloads 0 Views 2MB Size
Download PDF
Inorg. Chem. 2010, 49, 8953–8961 8953 DOI: 10.1021/ic1013472

cis-2,20 -Bipyrimidine-Bridged Polynuclear Complex: A Stairway-like Mixed-Valent {Fe4} Cluster Pablo Albores* and Eva Rentschler*,† Institut f€ ur Anorganische Chemie und Analytische Chemie, Johannes Gutenberg-Mainz Universit€ at, Duesbergweg 10-14, D-55099 Mainz, Germany. † E-mail: [email protected]. Tel: þ54 (011) 4576-3380. Fax: þ54 (011) 4576-3343. Received July 6, 2010

We report the first example of a polynuclear discrete coordination compound exhibiting only bpym bridges and containing a first-row d transition metal. A smooth self-assembly one-pot synthetic route, starting from simply FeCl2 and FeCl3 hydrates, allowed us to prepare a tetranuclear Fe4 cluster with a stairway-like structure and the formula cis-{[(H2O)Cl3FeIII-μ(bpym)FeIICl2]}2-μ(bpym) (1) . All spectroscopic data suggest that complex 1 is a valencelocalized mixed-valent FeII-FeIII cluster with typical M€ossbauer lines for both sites, which do not change with temperature. Reflectance spectroscopy did not allow one to distinguish an intervalence charge-transfer band. However, time-dependent density functional theory (DFT) calculations predict a weak high-energy FeII f FeIII transition. Regarding the magnetic properties, the high-spin FeII and FeIII ions interact in a weakly antiferromagnetic way with isotropic J constants of only a few wavenumbers as derived from direct-current susceptibility and magnetization data. Broken-symmetry DFT calculations support these observations.

Introduction 2,20 -Bipyrimidine (bpym) is the key member of the pyridine-type bis-bidentate ligands capable of bridging two different metal ions. In spite of its versatility to link a great variety of transition metals, there are to date only reports either of just dinuclear complexes1-4 or, on the contrary, of *To whom correspondence should be addressed. Current address: INQUIMAE-DQIAyQF, Universidad de Buenos Aires, Ciudad Universitaria, Pabellon 2, 3 piso, Buenos Aires C1428EHA, Argentina. E-mail: albores@ qi.fcen.uba.ar. Tel: þ49 (0)6131 39 25491. Fax: þ49 (0)6131 39 23922. (1) (a) Brewer, G.; Sinn, E. Inorg. Chem. 1985, 24, 4580. (b) Real, A.; Zarembowitch, J.; Kahn, O.; Solans, X. Inorg. Chem. 1987, 26, 2939. (c) Demunno, G.; Julve, M.; Lloret, F.; Derory, A. J. Chem. Soc., Dalton Trans. 1993, 1179. (d) Demunno, G.; Julve, M.; Verdaguer, M.; Bruno, G. Inorg. Chem. 1993, 32, 2215. (e) Demunno, G.; Julve, M.; Lloret, F.; Faus, J.; Caneschi, A. J. Chem. Soc., Dalton Trans. 1994, 1175. (f) Schoberl, U.; Magnera, T. F.; Harrison, R. M.; Fleischer, F.; Pflug, J. L.; Schwab, P. F. H.; Meng, X. S.; Lipiak, D.; Noll, B. C.; Allured, V. S.; Rudalevige, T.; Lee, S.; Michl, J. J. Am. Chem. Soc. 1997, 119, 3907. (g) Baumann, F.; Stange, A.; Kaim, W. Inorg. Chem. Commun. 1998, 1, 305. (h) Sieger, M.; Vogler, C.; Klein, A.; Knodler, A.; Wanner, M.; Fiedler, J.; Zalis, S.; Snoeck, T. L.; Kaim, W. Inorg. Chem. 2005, 44, 4637. (i) Gaspar, A. B.; Ksenofontov, V.; Reiman, S.; Gutlich, P.; Thompson, A. L.; Goeta, A. E.; Munoz, M. C.; Real, J. A. Chem.;Eur. J 2006, 12, 9289. (j) Govindaswamy, P.; Canivet, J.; Therrien, B.; Suss-Fink, G.; Stepnicka, P.; Ludvik, J. J. Org. Chem. 2007, 692, 3664. (k) Inagaki, A.; Yatsuda, S.; Edure, S.; Suzuki, A.; Takahashi, T.; Akita, M. Inorg. Chem. 2007, 46, 2432. (2) Demunno, G.; Ventura, W.; Viau, G.; Lloret, F.; Faus, J.; Julve, M. Inorg. Chem. 1998, 37, 1458. (3) Sun, J. S.; Zhao, H. H.; Xiang, O. Y.; Clerac, R.; Smith, J. A.; Clemente-Juan, J. M.; Gomez-Garcia, C.; Coronado, E.; Dunbar, K. R. Inorg. Chem. 1999, 38, 5841. (4) Thetiot, F.; Triki, S.; Pala, J. S.; Galan-Mascaros, J. R.; MartinezAgudo, J. M.; Dunbar, K. R. Eur. J. Inorg. Chem. 2004, 3783.

r 2010 American Chemical Society

one-dimensional (1D) infinite chains, with bpym being the unique bridging moiety.4-6 bpym has proven to mediate a moderate interaction between metal ions as evidenced from the electronic and magnetic properties of the dinuclear complexes. It also has been extensively used in combination with other bridging ligands in the design of two-dimensional (2D) and three-dimensional (3D) networks, mainly within the field of molecular-based magnets.4,7,8 The preparation of homodinuclear bpym-bridged compounds and the design of 1D coordination polymers appear to be quite reasonable tasks. Changing the stoichiometry of the metal fragments and bpym leads to the dinuclear species when the metal ions are used in excess, whereas just mixing one to one amounts of the metal source and bridging ligand leads to the polymeric form. However, the preparation of discrete clusters with intermediate nuclearity size bridged by bpym seems to be a real synthetic challenge for chemists. In fact, to the best of our (5) Pointillart, F.; Herson, P.; Boubekeur, K.; Train, C. Inorg. Chim. Acta 2008, 361, 373. (6) (a) Morgan, L. W.; Goodwin, K. V.; Pennington, W. T.; Petersen, J. D. Inorg. Chem. 1992, 31, 1103. (b) Demunno, G.; Viterbo, D.; Caneschi, A.; Lloret, F.; Julve, M. Inorg. Chem. 1994, 33, 1585. (c) Demunno, G.; Julve, M.; Real, J. A.; Lloret, F.; Scopelliti, R. Inorg. Chim. Acta 1996, 250, 81. (d) Cortes, R.; Urtiaga, M. K.; Lezama, L.; Pizarro, J. L.; Arriortua, M. I.; Rojo, T. Inorg. Chem. 1997, 36, 5016. (e) Demunno, G.; Poerio, T.; Julve, M.; Lloret, F.; Viau, G.; Caneschi, A. J. Chem. Soc., Dalton Trans. 1997, 601. (f) Demunno, G.; Poerio, T.; Julve, M.; Lloret, F.; Faus, J.; Caneschi, A. J. Chem. Soc., Dalton Trans. 1998, 1679. (g) Demunno, G.; Poerio, T.; Julve, M.; Lloret, F.; Viau, G. New J. Chem. 1998, 22, 299. (h) Armentano, D.; Demunno, G.; Guerra, F.; Faus, J.; Lloret, F.; Julve, M. Dalton Trans. 2003, 4626.

Published on Web 08/31/2010

pubs.acs.org/IC

8954 Inorganic Chemistry, Vol. 49, No. 19, 2010

Albores and Rentschler

Figure 1. ORTEP representation (50% probability) of complex 1. Hydrogen atoms were omitted for the sake of clarity. Symmetry-equivalent positions: -x, -y, -z.

knowledge, there is only one example reported to date of this type of molecule with full structural characterization, corresponding to a tetranuclear silver(I) compound of the formula [Ag4(hfac)4(μ2-bpym)3].5 No examples are available with first-row paramagnetic transition metals providing possibly interesting electronic and/or magnetic properties because of their variability of electronic and spin ground states. In this work, we report the first example of a polynuclear discrete coordination compound exhibiting only bpym bridges and containing a first-row d transition metal. A smooth selfassembly one-pot synthetic route starting from FeCl2 and FeCl3 hydrates allows us to prepare a tetranuclear Fe4 cluster, cis-{[(H2O)Cl3FeIII-μ(bpym)FeIICl2]}2-μ(bpym) (1), with a stairway-like structure. The mixed-valent FeII/FeIII character makes this novel compound even more interesting for chemists. We report herein its structural characterization and spectroscopic and magnetic properties, together with density functional theory (DFT) calculations, completing a deep study about this new bpym-bridged Fe4 cluster. Results and Discussion Synthetic Procedure. The reaction of equimolar amounts of FeCl3 and FeCl2 hydrates with bpym in a methanol/ acetonitrile mixture afforded single crystals of the mixedvalent cluster 1, after a first removal of an insoluble polycrystalline solid, which as not further characterized. The high insolubility in acetonitrile of 1 drives the selfassembly process, which most probably involves the (7) (a) Demunno, G.; Julve, M.; Lloret, F.; Faus, J.; Verdaguer, M.; Caneschi, A. Angew. Chem., Int. Ed. Engl. 1993, 32, 1046. (b) Demunno, G.; Julve, M.; Nicolo, F.; Lloret, F.; Faus, J.; Ruiz, R.; Sinn, E. Angew. Chem., Int. Ed. Engl. 1993, 32, 613. (c) Decurtins, S.; Schmalle, H. W.; Schneuwly, P.; Zheng, L. M.; Ensling, J.; Hauser, A. Inorg. Chem. 1995, 34, 5501. (d) Demunno, G.; Julve, M.; Viau, G.; Lloret, F.; Faus, J.; Viterbo, D. Angew. Chem., Int. Ed. Engl. 1996, 35, 1807. (e) Demunno, G.; Poerio, T.; Viau, G.; Julve, M.; Lloret, F.; Journaux, Y.; Riviere, E. Chem. Commun. 1996, 2587. (f) Armentano, D.; Demunno, G.; Lloret, F.; Julve, M. Inorg. Chem. 1999, 38, 3744. (g) Marshall, S. R.; Incarvito, C. D.; Manson, J. L.; Rheingold, A. L.; Miller, J. S. Inorg. Chem. 2000, 39, 1969. (h) Triki, S.; Thetiot, F.; Galan-Mascaros, J. R.; Pala, J. S. New J. Chem. 2001, 25, 954. (i) Armentano, D.; Demunno, G.; Lloret, F.; Julve, M.; Curely, J.; Babb, A. M.; Lu, J. Y. New J. Chem. 2003, 27, 161. (j) Colacio, E.; Debdoubi, A.; Kivekas, R.; Rodriguez, A. Eur. J. Inorg. Chem. 2005, 2860. (k) Colacio, E.; Lloret, F.; Navarrete, M.; Romerosa, A.; Stoeckli-Evans, H.; SuarezVarela, J. New J. Chem. 2005, 29, 1189. (l) Podgajny, R.; Pinkowicz, D.; Korzeniak, T.; Nitek, W.; Rams, M.; Sieklucka, B. Inorg. Chem. 2007, 46, 10416. (m) Baca, S. G.; Malaestean, I. L.; Keene, T. D.; Adams, H.; Ward, M. D.; Hauser, J.; Neels, A.; Decurtins, S. Inorg. Chem. 2008, 47, 11108. (n) Fabelo, O.; Pasan, J.; Lloret, F.; Julve, M.; Ruiz-Perez, C. Inorg. Chem. 2008, 47, 3568. (8) Albores, P.; Rentschler, E. Dalton Trans. 2009, 2609. (9) Oshio, H.; Hoshino, N.; Ito, T.; Nakano, M. J. Am. Chem. Soc. 2004, 126, 8805.

Figure 2. Atom-to-atom overlaying of crystal structure pairs: 1/2 (top) and 1/3 (bottom).

preformation of a previously characterized dinuclear compound cis-(bpym)Cl2FeII-μ(bpym)FeIICl2(bpym).3 Capping the latter with FeIIICl3 units and an additional water solvent molecule, which completes the octahedral environment, ends in the cis tetranuclear Fe4 species. This solid proves insoluble in all common solvents, with the exception of water, where it quickly decomposes. Structural Characterization. The X-ray structure of complex 1 reveals an all-cis configuration in an almost perfect stairway structure, with {FeIII-μ(bpym)-FeII} motifs (related by a crystallographic inversion center) lying as each stair step (Figure 1). Alternatively, by a 90° rotation, it can be described as a longer staircase, with each FeIII as the first and last stair steps and both FeII ions as the middle ones (see the Supporting Information, SI). Noticeably, as nice complements, structures are reported for the mononuclear complex [FeIIICl3H2O(bpym)]2 (2) and the dinuclear complex cis-[{FeII(bpym)Cl2}2-μ(bpym)]3 (3). The FeIII coordination sphere in 1 closely matches the observed one in complex 2, while the FeII backbone resembles the full structure of complex 3, as can be noticed from an atom-to-atom overlaying of both structures with rootmean-square (rms) values of 0.147 and 0.307 A˚, respectively (Figure 2). Overall, the whole structure of complex 1 can be viewed as a superposition of the crystal structures of complexes 2 and 3. In terms of its architecture, complex 1 is built up

Article

Inorganic Chemistry, Vol. 49, No. 19, 2010

Table 2. Selected Bond Lengths (A˚) and Bond Angles (deg) for Complex 1

Table 1. Crystal and Refinement Data 1 empirical formula fw temperature (K) wavelength (A˚) cryst syst space group unit cell dimensions

volume (A˚3) Z density calcd (g cm-3) abs coeff (mm-1) F(000) cryst size (mm3) θ range (deg) limiting indices reflns collected indep reflns completeness to θ (%) refinement method data/restraints/params GOF on F2 final R indices [I >2θ(Ι)] R indices (all data) largest diff peak and hole (e A˚-3)

8955

C24 H22 Cl10 Fe4 N12 O2. 2H2O 1124.47 173 0.710 69 monoclinic P21/n a = 7.4691(3) A˚ b = 11.7690(5) A˚ c = 22.9889(10) A˚ β = 96.8550(10)° 2006.4(2) 2 1.861 2.133 1120 0.12  0.11  0.05 1.78-27.45 -9 e h g 9 -15 e k g 15 -29 e l g 29 21 602 4563 [Rint =0.0389] 99.5 full-matrix least squares on F2 4563/4/256 1.091 R1 = 0.0345, wR2 = 0.0711 R1 = 0.0470, wR2 = 0.0781 0.497 and -0.407

from the coordination of neutral {FeIIICl3H2O} moieties to terminal bpym free sites of complex 2. As can be already realized from a comparison of the crystallographic data2,3 of complexes 2 and 3, in the case of complex 1, a small but sizable difference can be found in the Fe-N bond distances for both metal oxidation states (average FeIII-N distance, 2.241 A˚; average FeII-N distance, 2.192 A˚). The shorter FeII-N bond distance simply reflects the enhanced π-bonding ability of FeII versus FeIII ions. On the other hand, the Fe-Cl bond distances are clearly shorter for FeIII (average FeIII-Cl distance, 2.281 A˚; average FeII-Cl distance, 2.387 A˚), as was expected from the combination of the chloride π-donor ligand with the better acceptor ion FeIII in comparison with FeII (Table 2). The coordinated aqua ligand exhibits a considerably shorter FeIII-O bond distance than the one observed in complex 2, 2.059(2) vs 2.130 A˚, respectively. This is accompanied with a lengthening of the trans Fe-Cl distance [2.3230(8) vs 2.305 A˚] compared to complex 2. Regarding the local symmetry of both Fe sites, they appear considerably distorted from an octahedron, with angles of 73.56(8)° (N3-Fe1-N5) and 106.99(3)° (Cl3Fe2-Cl4). All of the metric data indicate localized FeII and FeIII sites within this mixed-valent compound crystal structure, in agreement with the marked superposition of the solid-state structures of the FeIII-only compound 2 and the FeII-only compound 3. In the crystal packing, the supramolecular organization of complex 1 is best described as adjacent tetranuclear clusters hold together by π-π stacking between the outer bpym rings (mean ring-ring distance ca. 3.94 A˚). Hydrogen bonding between coordinated water and the neighboring FeII-Cl4 (3.161 A˚ for O1 3 3 3 Cl4) completes the supramolecular arrangement (see the SI). Hence, with 6.708 A˚,

Fe1-Cl1 Fe1-Cl2 Fe1-Cl5 Fe1-O1 Fe1-N3 Fe1-N5 Fe2-Cl3 Fe2-Cl4 Fe2-N1 Fe2-N2 Fe2-N4 Fe2-N6 Fe1 3 3 3 Fe2 Fe2 3 3 3 Fe20

2.2618(8) 2.2588(8) 2.3230(8) 2.059(2) 2.220(2) 2.263(2) 2.3668(8) 2.4081(8) 2.166(2) 2.201(2) 2.175(2) 2.225(2) 5.882 5.831

N1-Fe2-N2 N4-Fe2-N2 N1-Fe2-N6 N4-Fe2-N6 N2-Fe2-N6 N1-Fe2-Cl3 N4-Fe2-Cl3

75.67(8) 101.04(8) 102.83(9) 74.99(8) 81.80(8) 91.50(6) 91.11(6)

N6-Fe2-Cl3 N1-Fe2-Cl4 N4-Fe2-Cl4 O1-Fe1-N3 N3-Fe1-Cl2 O1-Fe1-Cl2 O1-Fe1-Cl1 Cl2-Fe1-Cl1 O1-Fe1-N5 N3-Fe1- N5 Cl1-Fe1-N5 N3- Fe1-Cl5 Cl2-Fe1-Cl5 Cl1-Fe1-Cl5 N5-Fe1-Cl5 N2-Fe2-Cl4 Cl3-Fe2-Cl4 N2-Fe2-N6

82.72(6) 90.17(7) 91.50(6) 82.01(9) 93.41(6) 92.32(7) 90.96(7) 102.98(3) 86.88(8) 73.56(8) 90.08(6) 86.53(6) 94.95(3) 98.14(3) 83.54(6) 92.18(7) 106.99(3) 81.80(8)

Figure 3. Diffuse-reflectance spectrum of 1. Reference sample: barium sulfate.

the shortest intercluster Fe 3 3 3 Fe distance (Fe1neighbor1Fe2neighbor2) is considerably longer than the intracluster Fe-Fe distances (Fe1-Fe2 = 5.882 A˚ and Fe2-Fe20 = 5.831 A˚). These intermolecular interactions propagate along the structure, affording 2D sheets connected through a further hydrogen-bonding network involving the free water molecule, the coordinated aqua ligand, and coordinated Cl3 and Cl5 ligands (see the SI). An overall zigzag arrangement can be observed for the π-stacked sheets of Fe4 complexes. A similar pattern has been observed in the crystal packing of the related dinuclear complex, [{FeCl3(H2O)}2-μ(bpym)]2 (4). In this case, the intermolecular Fe 3 3 3 Fe distance is somewhat shorter, 6.422 A˚. Reflectance Spectrum. Because of the high insolubility of complex 1 in all common solvents, we studied its electronic spectroscopy in the solid state by measuring the spectrum in the diffuse-reflectance mode (Figure 3). The spectrum displays two broad intense bands in the visible region at 25.0  103 and 16.9  103 cm-1, attributable to Cl f FeIII ligand-to-metal charge transfer (LMCT) and FeII f bpym metal-to-ligand charge transfer (MLCT). In the related Fe(bpym)2(CN)2 compound, the FeII f bpym MLCT bands appear at 26.3 and 18.3  103 cm-1, while the Cl f FeIII LMCT bands have been

8956 Inorganic Chemistry, Vol. 49, No. 19, 2010

Albores and Rentschler

Table 3. Electronic Spectroscopic Data and TD-DFT Results exptl wavenumber/ 103 cm-1 (F(R)) 25.0 (0.25)

1

not resolved

16.9 (0.15)

8.3 (0.008) Fe(bpym)2Cl2

calcd wavenumber/103 cm-1 (osc str)

main CI determinant contributionsa

17.1 (0.0105) 17.5 (0.0109) 17.2 (0.0155) 16.5 (0.0124) 16.4 (0.0086) 14.5 (0.0047) 13.7 (0.0040) 13.5 (0.0010) 12.8 (0.0013) 11.9 (0.0057) 10.0 (0.0087)

H(β) f Lþ2(β), H-1(β) f L(β), Lþ2(β) H(β) f Lþ14(β), H-1(β) f Lþ13(β) H-6 (β) f Lþ1 (β), L (β), H-7 (β) f L (β) H-5(β) f Lþ6(β), H-9(β) f Lþ7(β) H(β) f Lþ12(β) H-1 (β) f Lþ3(β) H(β) f Lþ6(β), Lþ8(β), H-1(β) f Lþ7(β) H(β) f Lþ4(β), H-1(β) f Lþ5(β) H(β) f Lþ6(β), Lþ8(β), H-1(β) f Lþ3(β) H(β) f Lþ1(β), L(β), H-1(β) f L(β) H(β) f Lþ27(β), Lþ2(β), H-1(β) f Lþ28(β)

9.1 (0.0189) 1.6 (0.0027)

H(β) f L(β), Lþ1(β), H-1(β) f L(β) H(β) f Lþ19(β), Lþ22(β), Lþ2(β), H-1(β) f Lþ21(β), Lþ9 (β) H(β) f Lþ4(β), Lþ5(β) H(β) f L(β), Lþ6(β)

18.1 (0.0089) 11.0 (0.0138) 1.4 (0.0007)

Fe(bpym)Cl3(H2O) a

20.4 (0.0058)

H(β) f L(β), Lþ3(β), Lþ4(β), Lþ6(β), Lþ7(β) H-2(β) f Lþ1(β), H-3(β) f Lþ1(β), H-4 f Lþ2(β), H-7(β) f Lþ2(β)

assignment LMCT, π Cl f dπ FeIII MLCT, dπ FeII f π bpym LL0 CT, π Cl f π bpym LMCT, π Cl f dπ FeIII MLCT, dπ FeII f π bpym MMCT, dπ FeII f dπ FeIII

MLCT, dπ FeII f π bpym dπ FeII f dπ FeII MLCT, dπ FeII f π bpym MLCT, dπ FeII f π bpym dπ FeII f dπ FeII MLCT, dπ FeII f π bpym MLCT, dπ FeII f π bpym dπ FeII f dπ FeII dπ FeII f dπ FeII MLCT, dπ FeII f π bpym LMCT, π Cl f dπ FeIII

|CI| coefficients >0.3.

Figure 4. TD-DFT-calculated spectra, at the B3LYP level with a LanL2DZ basis set in vacuum, of complex 1 and its constituent fragments Fe(bpym)2Cl2 and Fe(bpym)Cl3H2O. The wavenumber range of calculated spectra is shown to shift with respect to the experimental one according to the linear correlation found (Figure S5 in the SI) between them.

observed in the range (40.0-25.0)  103 cm-1 for several chloride ferric species.10,11 An additional much less intense band can be observed in the near-IR (NIR) region at 8.3  103 cm-1 with a visible shoulder at the high energy side (ca. 10.0  103 cm-1). The latter features are commonly observed in a high-spin d6 FeII ion at this energy region and correspond to the low-symmetry split 5Eg f 5 T2g d-d unique spin-allowed transition.11 For example, they are observed at 8.5 and 10.5  103 cm-1 in the related complex cis-Fe(phen)2Cl2.11 Because of the impossibility to perform a spectroelectrochemistry experiment in (10) Ruminski, R. R.; Vantassel, K. D.; Petersen, J. D. Inorg. Chem. 1984, 23, 4380. (11) Lever, A. B. P. Inorganic Electronic Spectroscopy, 2nd ed.; Elsevier Science Publishers BV: Amsterdam, The Netherlands, 1984.

solution, we decided to perform time-dependent DFT (TD-DFT) calculations to support the electronic spectral band assignments and explore the lack of an intervalence FeII f FeIII metal-to-metal charge-transfer (MMCT) band expected in this type of mixed-valent compound.12 The calculated electronic spectrum of 1 (Figure 3 and Table 3) is quite convincing and supports the presence of an intervalence charge-transfer (IVCT) active transition in the spectrum but buried under the LMCT and MLCT bands because of its low intensity. All of the expected band assignments are supported by the calculations, with predominant LMCT, MLCT, and FeII d-d bands (Figure 4). The additional calculated electronic spectra of the isolated neutral moieties cisFe(bpym)2Cl2 and Fe(bpym)Cl3(H2O) show that the complex 1 spectral profile closely resembles the rough summation of the individual spectral contributions of both building fragments. The band profiles and relative intensities of the calculated spectrum nicely reproduce the experimental one, differing only in a fixed factor (ascribable to intrinsic DFT limitations), as demonstrated by the tight linear correlation (with a unitary slope) between the experimental and calculated transition energies (see the SI). The apparent low intensity of the unobserved FeII f FeIII MMCT seems to originate in the low overlap between the dπ donor and acceptor orbitals set as inferred from the molecular orbitals (MOs) involved in this transition calculated by TD-DFT (Figure 5). The cis arrangement of the Fe(bpym)2Cl2 moieties might be responsible for this symmetry mismatching. M€ossbauer Spectrosocopy. Zero-field M€ ossbauer spectra of complex 1 were measured at 80, 180, and 297 K (Figure 6). The spectral parameters obtained from a leastsquares fit to the experimental points, assuming Lorentzian absorption lines, are summarized in Table 4. (12) D’Alessandro, D. M.; Keene, F. R. Chem. Rev. 2006, 106, 2270.

Article

Inorganic Chemistry, Vol. 49, No. 19, 2010

8957

Figure 6. M€ ossbauer spectra at different temperatures of complex 1. Full line: fitted convoluted spectra. Dashed lines: fitted Lorentzian doublets for each Fe site. Table 4. M€ossbauer Spectra Fitted Parameters site I

site II

δ ΔEQ Γ/2 δ ΔEQ Γ/2 T/K [mm s-1] [mm s-1] [mm s-1] [mm s-1] [mm s-1] [mm s-1] 80 180 297

Figure 5. Main CI MOs contributing to the calculated MMCT transition. Arrows indicate the more intense contributions.

The spectrum of 1 at 80 K consists of two, 1:1 integrated area, nested quadrupole split doublets consistent with two different Fe sites and the centrosymmetric structure of 1 (also supported by the rather small line-width values). The isomer shifts of the two doublets are 0.49 and 1.14 mm s-1, in the range expected for six-coordinated highspin FeIII and FeII ions, respectively,13 and in agreement with a localized mixed-valence species description. The distinctly different quadrupole splittings, ΔEQ, of 0.29 and 2.56 mm s-1 are in the range observed for high-spin d5 FeIII ions and high-spin d6 FeII ions in comparable environments. For example, ΔEQ values observed for FeCl3, FeCl3 3 6H2O, FeCl2 3 H2O, and Fe(phen)2Cl2 are 0, 0.97, 2.03, and 3.28 mm s-1, respectively.13 With increasing temperature, the two-doublet feature is retained and the line widths almost keep the same small values. The smooth decrease observed for both Fe sites in the isomer shifts, with increasing temperature, is expected (13) Gutlich, P.; Ensling, J. In Inorganic Electronic Structure and Spectroscopy; Solomon, E. I., Lever, A. B. P., Eds.; Wiley-VCH: New York, 1999; Vol. 1.

0.49 0.46 0.37

0.29 0.31 0.29

0.19 0.15 0.22

1.14 1.08 1.00

2.56 2.18 1.92

0.21 0.18 0.24

for a second-order Doppler shift.14 Regarding the other key parameter ΔEQ, FeIII site quadrupole splitting is unchanged with increasing temperature, while the ΔEQ value for the FeII doublet decreases from 2.56 to 1.92 mm s-1 upon heating to 297 K. This shifting is typical for highspin ferrous systems because of changes in the Boltzmann populations of the low-lying electronic states that form as a consequence of spin-orbit and low-symmetry crystalfield perturbations.15 Overall, a mixed-valent localized picture for this Fe4 compound is suggested from M€ ossbauer data up to room temperature. Magnetic Properties. The direct-current (dc) magnetic susceptibility of complex 1 was measured in the temperature range 2-300 K (Figure 7) under an applied field of 1 T. At 300 K, a χmT (χm = molar susceptibility) value of 14.61 cm3 K mol-1 is observed, quite close to the expected value for two isolated high-spin FeIII ions (S = 5/2) and two isolated high-spin FeII ions (S = 2), 14.76 cm3 K mol-1 with g = 2. With decreasing temperature, the χmT product continuously decreases, approaching a zero value, and finally reaches a value of 0.89 cm3 K mol-1 at 2 K. This behavior clearly indicates that antiferromagnetic interactions (weak ones as inferrd from the χmT value at 300 K) are operative between the Fe centers. This magnetic behavior has been previously observed in tightly related bipyrimidine-bridged Fe complexes, as in the (14) Lynch, M. W.; Valentine, M.; Hendrickson, D. N. J. Am. Chem. Soc. 1982, 104, 6982. (15) (a) Reiff, W. M.; Dockum, B.; Weber, M. A.; Frankel, R. B. Inorg. Chem. 1975, 14, 800. (b) Stassinopoulos, A.; Schulte, G.; Papaefthymiou, G. C.; Caradonna, J. P. J. Am. Chem. Soc. 1991, 113, 8686.

8958 Inorganic Chemistry, Vol. 49, No. 19, 2010

Albores and Rentschler

Figure 7. χmT and χm vs T plot at 1 T in the range 2-300 K of complex 1. Full line: best fitting with the Hamiltonian of eq 1. Dashed line: best fitting with the Hamiltonian of eq 1 and zfs Hamiltonian of eq 2 (see the text). Scheme 1. Exchange Coupling Pattern Describing Magnetic Interactions in Complex 1

already mentioned dinuclear FeII complex 3, [{FeII(bpym)Cl2}2-μ(bpym)], and the dinuclear FeIII compound 4, [{FeCl3(H2O)}2-μ(bpym)].2,3 We attempted a full fitting of the data employing the MAGPACK package16, where the energy of the different spin levels is obtained through diagonalization of the suitable Hamiltonian. In this case, the Heisenberg spin Hamiltonian (with the corresponding Zeeman terms) describing the isotropic exchange interactions within this Fe4 complex (Scheme 1) is given by eq 2, where J1 refers to the interactions between the FeIII-FeII pairs and J2 refers to the FeII-FeII interaction: ^ ¼ - 2J1 ðS^Fe1 S^Fe2 þ S^Fe10 S^Fe20 Þ - 2J2 ðS^Fe2 S^Fe20 Þ H 3 3 3 ð1Þ Both FeIII-FeII exchange interactions must be identical because of the crystallographically imposed inversion center. A satisfactory fitting of the χmT vs T experimental data in the whole temperature range was obtained (Figure 7) with the following parameters: gav = 2.06 ( 0.01, J1 = -2.0 ( 0.6 cm-1, and J2 = -1.9 ( 0.6 cm-1 (R = 3.32  10-4). These values for the exchange coupling constants J Fe-Fe are in agreement with the previously found values in complexes 3 (JFeII-FeII = -1.1 cm-1) and 4 (JFeIII-FeIII = -0.4 cm-1).2,3 It must be noticed at this point that an indistinguishable fit of the data can be obtained by applying the restraint J1 = J2, with parameters gav = 2.06 ( 0.01 and J1 = -1.99 ( 0.05 cm-1. This can be better understood as an (16) Borras-Almenar, J. J.; Clemente-Juan, J. M.; Coronado, E.; Tsukerblat, B. S. J. Comput. Chem. 2001, 22, 985.

Figure 8. Μ vs H/T plot of complex 1 in the ranges 2-5 K and 10-70 kOe. The lines correspond to simulated and/or fitted plots (see the text): gray full line, Hamiltonian eq 1 with fixed χmT vs T best-fitting parameters; gray dashed line, Hamiltonian eq 1 with a fixed gav value; black dashed line, Hamiltonian eq 1 and zfs eq 2 with fixed χmT vs T best-fitting parameters; black full line, Hamiltonian eq 1 and zfs eq 2 with fixed χmT vs T best-fitting J1 and J2 parameters.

equivalence of both J values rather than a sort of overparameterization, as can be seen in a J1-J2 correlation error surface (see the SI). However, below 15 K, the fit of the data is poor and the simulated data underestimate the experimental data. This discrepancy cannot be attributed to intermolecular antiferromagnetic exchange between adjacent Fe4 clusters, mediated by hydrogen bonding (see the Structural Characterization section), as previously observed in the related dinuclear compound 42 because this interaction should lower the χm data at low T. The most reasonable explanation must rely on the well-known single-ion zerofield-splitting (zfs) contribution of the high-spin d6 FeII sites. In fact, this contribution is observed in the susceptibility data of the dinuclear complex 3, [{FeII(bpym)Cl2}2-μ(bpym)], where a negative D value of about 17 cm-1 must be included to properly account for the experimentally observed data at low temperature.3 When the proper zfs Hamiltonian term (eq 2) is added to eq 1, the best data-fitting parameters become gav=2.05 ( 0.01, J1 = -2.2 ( 0.2 cm-1, J2 = -0.6 ( 0.2 cm-1, and D = 17 ( 1 cm-1 (R = 1.14  10-4). ^ zfs ¼ Df½S^2 - SFe1 ðSFe1 þ 1Þ=3 H Fe1 þ ½S^Fe10 - SFe10 ðSFe10 þ 1Þ=3g 2

ð2Þ

The low-temperature χm data are now nicely reproduced (Figure 7), while only the FeII-FeII J2 exchange coupling constant value is somewhat lower than the value obtained with complete neglect of the zfs term. In spite of certain overparameterization risks, no strong correlation is observed between J1, J2, and D parameters, as can be observed in the respective contour error plots (see the SI). From these same plots, it can also be noticed that the alternative set of parameters with a negative D value affords local minima with much poorer goodness of fit values. From both negative exchange coupling constant values, a diamagnetic S = 0 ground state is established. However, low-energy-lying magnetic states are expected because of the weak magnitude of both J parameters. Precisely, when the zfs effects on the energy level diagrams are omitted (see

Article

the SI), the first excited state is S = 1 at only 1.3 cm-1 from the ground state, followed by S = 2 at 5.7 cm-1. In order to obtain more information about the lowlying energy levels and of the single-ion zfs contribution of FeII sites, we performed magnetization measurements in the range 2-5 K under external magnetic fields up to 70 kOe (Figure 8). No superimposed isofield magnetization curves are observed. The maximum reduced magnetization value reached at 2 K and 70 kOe is 5.44 Nβ, close to the expected saturation value for a pure spin S = 3 with a g value of around 2.0. This M vs H/T profile is in agreement with a diamagnetic ground state with lowlying magnetic excited states, allowing energy level crossing. In fact, when the energy levels plot is inspected (see the SI), it is found that two crossings are probable around 15 kOe (S = 1) and 45 kOe (S = 2). At the maximum scanned field of 70 kOe, an additional excited state with S = 3 appears as the first low-lying excited state very close in energy to that of the ground state. This ground-state progression from S = 0 to almost S = 3 is in agreement with the observed progression of the saturation values of the increasing field individual isofield lines. Of course, at this point, we completely neglect further energy level splittings due to zfs contribution, which could certainly modify this description, but the present model allows a qualitative understanding of the reduced magnetization plot profiles. Modeling of the magnetization curves employing MAGPACK, with eq 2 Hamiltonian, and no zfs contribution is poor (Figure 8), even when independent J1 and J2 values are fitted. Fixing gav = 2.06 best values of J1 = -1.9 ( 0.8 cm-1 and J2 = -1.6 ( 0.9 cm-1 affords an agreement factor R = 1.5  10-3 but a poor low-field description. A clear improvement is achieved if the zfs contribution is included (eq 2), employing best-fitting parameters obtained from χmT vs T data (R = 1.2  10-3) or fixing just J1 and J2 and fitting the gav and D parameters. In the latter case, optimized parameters found are gav = 2.08 ( 0.02 and D = 12 ( 1 cm-1 (R = 1.2  10-3). Magnetization data support again a positive D value for the local zfs FeII contribution. It seems at least striking that just the opposite sign for the D value has been determined for the structurally related complex 3, [{FeII(bpym)Cl2}2-μ(bpym)].3 Only a few examples of unequivocal determinations of the D sign in high-spin FeII systems are reported, and both possible alternatives have been found.17,18 A similar positive D value was found for a related cis-bis(thiazoline)iron(II) complex.18 In the case of complex 1, the local pseudosymmetry of the FeII sites is roughly C2, and the splitting of ground state 5 T2g results in three orbital nondegenerate new quintuplets: 5B, 5B, and 5A. A further spin-orbit coupling interaction determined these quintuplets as zfs. The relative ligand-field strengths of the FeII coordination sphere (angle- and distance-dependent) conclude the final D sign as a first approximation. More experimental data (17) (a) Knapp, M. J.; Krzystek, J.; Brunel, L. C.; Hendrickson, D. N. Inorg. Chem. 2000, 39, 281. (b) Carver, G.; Tregenna-Piggott, P. L. W.; Barra, A. L.; Neels, A.; Stride, J. A. Inorg. Chem. 2003, 42, 5771. (c) Telser, J.; van Slageren, J.; Vongtragool, S.; Dressel, M.; Reiff, W. M.; Zvyagin, S. A.; Ozarowski, A.; Krzystek, J. Magn. Reson. Chem. 2005, 43, S130. (18) Ozarowski, A.; Zvyagin, S. A.; Reiff, W. M.; Telser, J.; Brunel, L. C.; Krzystek, J. J. Am. Chem. Soc. 2004, 126, 6574.

Inorganic Chemistry, Vol. 49, No. 19, 2010

8959

Table 5. DFT-Calculated Exchange Coupling Constants J for Complex 1 Ruiz formalism (Ising formalism)

LanL2DZ

TZVP

J1/cm-1 -1.56 (-1.87)

-1.62 (-1.94)

J2/cm-1 -1.59 (-1.98)

-1.83 (-2.30)

experimental

χmT data

M vs H/ T data

-2.0 ( 0.6 -1.9 ( 0.8 -2.2 ( 0.2 (zfs) -1.9 ( 0.6 -1.6 ( 0.9 -0.6 ( 0.2 (zfs)

Figure 9. Magnetic orbitals after a COT describing the J1 and J2 exchange interactions.

(high-field electron paramagnetic resonance) or an elaborated theoretical approach must be performed to get a final definitive assessment regarding the sign and magnitude of the FeII zfs contribution.9,19 Complementary DFT broken-symmetry calculations were performed to obtain theoretical exchange coupling constant values and to inspect the nature of the exchange pathways in terms of magnetic orbital descriptions. The J1 and J2 calculated values with two different basis sets and the two different spin formalism approaches employed are listed in Table 5. It can be noticed that very close values are obtained for J1 and J2, both in good agreement with the experimental findings, independent of the method or basis set used. Simulated χmT vs T plots employing these theoretical values show excellent results (see the SI). These theoretical calculations support the weak antiferromagnetic interactions between Fe centers, as was expected for the bpym ligand. The nonzero overlapping magnetic orbitals for the FeII and FeIII sites are shown in Figure 9, represented through the corresponding orbital transformation. Both broken-symmetry topology contributions have been merged into a unique set of R-β pairs. As was also verified in some previous bpym-bridged examples, the main operative exchange pathways rely on pure σ-type orbital overlaps between the metal centers (19) Krzystek, J.; Ozarowski, A.; Telser, J. Coord. Chem. Rev. 2006, 250, 2308.

8960 Inorganic Chemistry, Vol. 49, No. 19, 2010 and the bpym ligand.8,20 Because of the local low symmetry, the d orbitals involved are not the expected dσ and dπ sets but just combinations of them. The similar overlap integrals describing FeIII-FeII and FeII-FeII exchange interactions justified the close values obtained for J1 and J2 isotropic constants. Conclusion We have successfully prepared the first example of a polynuclear discrete coordination compound exhibiting only bpym bridges and containing a first-row d transition metal. All spectroscopic data show that this is a valence-localized ossbauer lines mixed-valent FeII-FeIII cluster with typical M€ for both sites and not an experimental observed but a theoretically predicted weak high-energy IVCT transition. Regarding the magnetic properties, the high-spin FeII and FeIII ions interact in a weakly antiferromagnetic way with isotropic J constants of a few wavenumbers. To our knowledge, this is just the second example of a structurally characterized discrete polynuclear complex with all bpym bridges and the first mixed-valent one. In this sense, the complex 1 structure becomes a remarkable breakthrough in the field of polynuclear transition-metal cluster compounds, and it may become a milestone for these types of systems. The smooth self-assembly one-pot synthetic route leaves the door open for further exploration of these types of complexes, which occupy the vacant place between dinuclear and coordination polymer bpym-bridged systems. Experimental Section Material and Physical Measurements. The ligand 2,20 -bipyrimidine (bpym) was prepared following a previously reported procedure.21 All other chemicals were reagent grade and were used as received without further purification. Elemental analysis for C, H, and N was performed on a Foss Heraeus Vario EL elemental analyzer. UV-vis-NIR spectra were recorded with a Cary 5E spectrophotometer using the diffuse-reflectance technique with barium sulfate as a reference. Magnetic measurements were performed with a Quantum Design MPMS XL SQUID magnetometer. dc measurements were conducted from 2 to 300 K at 1 T and from 2 to 5 K under an applied field up to 7 T. All measurements were performed on restrained polycrystalline samples in order to avoid field-induced reorientation of the microcrystals. Experimental magnetic data were corrected for diamagnetism of the sample holders and of the constituent atoms (Pascal’s tables). M€ ossbauer data were recorded on an alternating-current constant-acceleration spectrometer. The minimum experimental line width was 0.3 mm s-1 (full width at half-height). The sample temperature was maintained constant in an Oxford Instrument. Isomer shifts are quoted relative to iron metal at 300 K. Synthesis of the Complex cis-[Cl 3 (H 2 O)Fe III (bpym)]2μ(bpym)FeIICl2(bpym)]2 3 H2O (1). A total of 0.35 g (1.3 mmol) of FeCl3 3 6H2O and 0.26 g (1.3 mmol) of FeCl2 3 4H2O were dissolved jointly in 30 mL of methanol to afford a clear orange solution. Separately, 0.31 g (2.0 mmol) of bpym was dissolved in 20 mL of acetonitrile. Both solutions were mixed, and a new dark-red solution was obtained. After 30 min, it was filtered to remove a dark insoluble (not further characterized) solid, and the resulting clear solution was left undisturbed to slowly evaporate at room temperature. After a few days, dark-red block crystals of 1, suitable for X-ray measurement, were (20) (a) Julve, M.; Demunno, G.; Bruno, G.; Verdaguer, M. Inorg. Chem. 1988, 27, 3160. (b) Albores, P.; Rentschler, E. Dalton. Trans. 2010, 39, 5005. (21) Vlad, G.; Horvath, I. T. J. Org. Chem. 2002, 67, 6550.

Albores and Rentschler obtained. One specimen was collected for diffraction experiments, and the remaining material was collected by filtration and air-dried. The dried solid had only one solvation water molecule, in contrast with crystals, which have two crystallization water molecules. Yield: 0.138 g (20%). Anal. Calcd for C24H24Cl10Fe4N12O3: C, 26.05; H, 2.19; N, 15.19. Found: C, 26.05; H, 2.66; N, 15.32. X-ray Structure Determination. Crystals suitable for X-ray diffraction were obtained directly from the synthetic procedure and mounted in a glass fiber. The crystal structure was determined with a Bruker Smart APEX II CCD area detector diffractometer using graphite-monochromated Mo KR radiation (λ=0.710 73 A˚) at 173 K. Data were corrected for absorption with SADABS.22 The structures were solved by direct methods with SHELXS-9723 and refined by full-matrix least squares on F2 with anisotropic displacement parameters for non-hydrogen atoms with SHELXL-97.23 Hydrogen atoms were added geometrically and refined as riding atoms with a uniform value of Uiso, with the exception of hydrogen atoms of coordinated and free water molecules that were located in the difference map. Crystallographic and refinement data are shown in Table 1. CCDC 776615 contains the supplementary crystallographic data for this paper. These data can be obtained free of charge from the Cambridge Crystallographic Data Centre via www. ccdc.cam.ac.uk/data_request/cif. DFT Quantum Computations. DFT spin-unrestricted calculations were performed at the X-ray geometry using the Gaussian03 package (revision D.01)24 at the B3LYP level employing the medium-size LanL2DZ basis set and the more complete TZVP one. Tightly converged (10-8 Eh in energy) single-point calculations were performed in order to analyze the exchange coupling between the metallic ion centers. The methodology applied here relies on the broken-symmetry formalism, originally developed by Noodleman for self-consistent-field methods,25 which involves a variational treatment within the restrictions of a single spin-unrestricted Slater determinant built upon using different orbitals for different spins. This approach has been later applied within the framework of DFT.26 The high-spin (HS) and broken-symmetry (BS) energies were then combined to estimate the exchange coupling parameter J involved in the widely used Heisenberg-Dirac-van Vleck (HDvV) Hamiltonian.27 We have calculated the different spin topologies of broken-symmetry nature (see the SI) by alternately flipping the spin on the different metal sites (two different BS topologies in this case). The exchange coupling constants Ji can be obtained after the individual pairlike component spin interactions involved in the description of the different broken-symmetry states are considered. We used the two main methodologies reported: the Ising approach,28 in which the broken-symmetry states were directly considered as eigenstates of the HDvV Hamiltonian with the corresponding equation EBS - EHS ¼ 2J12 ð2S1 S2 Þ and the method proposed by Ruiz and co-workers,29 in which the following equation was applied: EBS - EHS ¼ 2J12 ð2S1 S2 þ S2 Þ, with S 2 < S 1 (22) Sheldrick, G. M. SADABS, Multiscan Absorption Correction Program; University of G€ottingen: G€ottingen, Germany, 1996. (23) Sheldrick, G. M. SHELXS-97 and SHELXL-97, Programs for Crystal Structure Resolution; University of G€ottingen: G€ottingen, Germany, 1997. (24) Frisch, M. J.; et al. Gaussian03, revision D.01; Gaussian Inc.: Pittsburgh, PA, 2003. (25) Noodleman, L. J. Chem. Phys. 1981, 74, 5737. (26) Noodleman, L.; Baerends, E. J. J. Am. Chem. Soc. 1984, 106, 2316. (27) Kahn, O. Molecular Magnetism; VCH: New York, 1993. (28) Dai, D. D.; Whangbo, M. H. J. Chem. Phys. 2003, 118, 29. (29) Ruiz, E.; Rodriguez-Fortea, A.; Cano, J.; Alvarez, S.; Alemany, P. J. Comput. Chem. 2003, 24, 982.

Article In both cases, a set of linear equations must be solved to obtain the J parameters. Additionally, we have also employed the BS-type spin-unrestricted solutions after a corresponding orbital transformation (COT)30 as a means to visualize the interacting nonorthogonal magnetic orbitals and, hence, the spin-coupling exchange pathways. These orbitals do not have a well-defined orbital energy; for this reason, orbital energies are not given explicitly but only their overlapping magnitudes. The TD-DFT computation at the X-ray geometry only with the medium-size LanL2DZ basis set was calculated for the HS state to assist in the interpretation and assignment of the electronic spectrum. The output contained information for the excited-state energies, oscillator strengths (f), and a list of transitions that give rise to each excited state. The orbitals involved as well as the orbital contribution coefficients of the transitions were obtained. The molar absorptivity was calculated with a full width at half-maximum of 3000 cm-1. As an additional tool, we performed identical TD-DFT calculations for the complex 1 constituents, neutral isolated fragments FeIII(H2O)Cl3bpym and cis-FeII(bpym)2Cl2, with their geometries arising from the X-ray structure of the Fe4 complex. (30) Neese, F. J. Phys. Chem. Solids 2004, 65, 781.

Inorganic Chemistry, Vol. 49, No. 19, 2010

8961

Acknowledgment. We gratefully acknowledge the Alexander von Humboldt Foundation for granting a postdoctoral fellowship and Bernd Mienert for M€ ossbauer measurements. This work was partially supported by the National Center for Supercomputing Applications under Grant TG-MCA05S010. P.A. is a member staff of CONICET. Supporting Information Available: ORTEP plot (50% probability) of molecular structure of complex 1, showing an alternative staircase topology (Figure S1), ball-and-stick representation of the complex 1 crystal structure unit cell (Figure S2), ball-and-stick representation of intermolecular interactions in the complex 1 crystal structure (Figures S3 and S4), TD-DFT-calculated energy vs experimental transition correlation plot (Figure S5), J1-J2 error contour plots corresponding to the χmT vs T simulation (Figures S6 and S7), J1-D error contour plot corresponding to the χmT vs T simulation (Figures S8 and S9), low-lying energy level plot (Figure S10), χmT vs T simulation with DFT-calculated J values (Figure S11), spin-density contours of the nonequivalent broken-symmetry spin states (Figure S12), and X-ray crystallographic file of complex 1 in CIF format. This material is available free of charge via the Internet at http://pubs.acs.org.